Hi nyikile ti terabytes ta ti log ta CI eka LLM
Mavonelo
Mewayz Team
Editorial Team
Mugodi wa Nsuku lowu Fihliweke lowu Tshameke eka Phayiphi ya Wena ya CI
Xipano xin'wana na xin'wana xa vunjhiniyara xi ti tumbuluxa. Millions of lines, every single day — timestamps, stack traces, dependency resolutions, test results, build artifacts, and cryptic error messages that scroll past faster than anyone can read. CI logs are the exhaust fumes of modern software development, and for most organizations, they're treated exactly like exhaust: vented into storage and forgotten. But what if those logs contained patterns that could predict failures before they happen, identify bottlenecks costing your team hundreds of hours per quarter, and reveal systemic issues that no single engineer ever sees? We decided to find out by feeding terabytes of CI log data into a large language model — and what we discovered changed how we think about DevOps entirely.
Ha yini Tilog ta CI ti ri Data leyi nga tirhisiwiki ngopfu eka Vuinjhiniyere bya Software
Xiya vholumo ya sheer. Xipano xa vunjhiniyara xa le xikarhi lexi fambisaka 200 wa ku aka hi siku eka vuhlayiselo byo tala xi humesa kwalomu ka 2-4 GB ya datha ya raw log siku na siku. Ku tlula lembe, sweswo swi tlula terabyte ya tsalwa leri hlelekeke na leri nga hlelekangiki ngopfu leri khomaka nhlengeleto wun’wana na wun’wana, ku hetisisiwa kun’wana na kun’wana ka suite ya xikambelo, goza rin’wana na rin’wana ro tirhisa, na movha wun’wana na wun’wana wo tsandzeka lowu sisiteme ya wena yi tshameke yi hlangana na wona. I rhekhodo leyi heleleke ya vuyimburi ya vuhumelerisi bya nhlangano wa wena wa vunjhiniyara — naswona kwalomu ka ku hava munhu loyi a yi hlayaka.
Xiphiqo a hi leswaku data yi pfumala nkoka. Hi leswaku ratio ya signal-to-noise i brutal. Ku tsutsuma ka CI loku tolovelekeke ku humesa magidi ya milayeni ya vuhumelerisi, naswona kumbexana 3-5 wa milayeni yoleyo yi na vuxokoxoko lebyi nga tekiwaka goza. Vanjhiniyara va dyondza ku skena ku kuma tsalwa ro tshwuka, ku grep ku kuma "FAILED," ivi va ya emahlweni. But the patterns that matter most — the flaky test that fails every Tuesday, the dependency that adds 40 seconds to every build, the memory leak that only surfaces when three specific services run concurrently — those patterns are invisible at the individual log level. Swi humelela ntsena hi xikalo.
Traditional log analysis tools like ELK stacks and Datadog can aggregate metrics and surface keyword matches, but they struggle with the semantic complexity of CI output. A build failure message that reads "connection refused on port 5432" and one that reads "FATAL: password authentication failed for user 'deploy'" are both database-related failures, but they have completely different root causes and solutions. Ku twisisa ku hambana koloko swi lava muxaka wa ku anakanya hi mongo loku, ku fikela sweswinyana, vanhu ntsena a va ta ku nyika.
Xikambelo: Ku phamela 3.2 Terabytes ya Matimu yo Aka eka LLM
Setup a yi straightforward hi concept na nightmarish eka execution. Hi hlengeletile tin’hweti ta 14 ta tilog ta CI ku suka eka pulatifomo leyi tirhelaka vatirhisi vo tlula 138,000 — ku katsa ku aka eka vukorhokeri byo tala, tindhawu, na swikongomelo swa ku tirhisiwa. Dataset ya raw yi fike eka 3.2 terabytes: kwalomu ka 847 wa timiliyoni ta milayeni ya log ya munhu hi xiyexe leyi hlanganisaka 1.6 wa timiliyoni ta ku famba ka tiphayiphi ta CI. Hi chunked, embedded, na indexed data leyi, kutani hi aka retrieval-augmented generation (RAG) pipeline leyi nga hlamulaka swivutiso swa ririmi ra ntumbuluko mayelana na matimu ya hina yo aka.
Ntlhontlho wo sungula a ku ri ku lulamisa ka ha ri emahlweni. Tilog ta CI a hi tsalwa leri tengeke. They contain ANSI color codes, progress bars that overwrite themselves, binary artifact checksums, and timestamps in at least four different formats depending on which tool generated them. We spent three weeks just on normalization — stripping noise, standardizing timestamps, and tagging each log segment with metadata about which pipeline stage, repository, branch, and environment it belonged to.
Ntlhontlho wa vumbirhi a ku ri ku durha. Ku tsutsuma inference eka ti terabytes ta tsalwa a swi chipa, hambi ku ri hi aggressive chunking na retrieval optimization. Hi hisile hi ku tirhisa swikweleti swa nkoka swa xibalo eka n’hweti yo sungula ntsena, ngopfungopfu hikuva endlelo ra hina ro sungula a ri ri ra vuphukuphuku ngopfu — ku rhumela mongo wo tala ngopfu hi xivutiso na ku nga hlawuli hi laha ku ringaneke mayelana na leswaku hi swihi swiphemu swa log leswi faneleke. By the end of the second month, we'd reduced per-query costs by 87% through better embedding strategies and a two-stage retrieval system that used a smaller model to pre-filter before sending to the larger one.
Tiphetheni ta Ntlhanu leti LLM yi Ti Kumeke Leti Vanhu A va nge Pfuki va ti endlile
Eka vhiki ro sungula ro fambisa swivutiso, sisiteme yi humese vutivi lebyi a byi ta teka muhlahluvi wa munhu tinhweti ku byi kuma hi voko. Leswi a ku nga ri timhaka ta le tlhelo kumbe ku lava ku tiva — a ku ri timhaka ta mafambiselo leti humesaka ngati ya tiawara ta xiviri ta vunjhiniyara.
- The phantom dependency cascade. Ku pfuxetiwa ka phasela rin’we ra npm tin’hweti ta 9 emahlweni ka sweswo a ku nghenise ku hlwela ka 22 wa tisekoni eka ku aka kun’wana na kun’wana ka JavaScript. Ku hlwela ku fihliwile hikuva ku fambisana na ku ndlandlamuxiwa ka switirhisiwa swa CI leswi endleke leswaku ku aka ku hatlisa hi ku angarhela. Net-net, ti builds ti vonaka ti hatlisa, kambe tingava ti hatlisa 22 seconds still. Ku tsemakanya 400+ wa ku aka ka JS hi siku, sweswo a ku ri 2.4 wa tiawara ta xibalo lexi tlangisiwaka siku na siku.
- The timezone flake. Suite ya swikambelo a yi ri na mpimo wa ku tsandzeka wa 4.7% — ntsena ehenhla swinene ku va yi hlundzukisa, ntsena ehansi swinene lerova a ku na munhu loyi a rhangisaka emahlweni ku yi lulamisa. LLM yi kumile leswaku ku tsandzeka ku fambelana kwalomu ka hi ku hetiseka na ku aka loku hlohloteriweke exikarhi ka 23:00 na 01:00 UTC, loko ntirho wo pimanisa siku wu tsemakanya ndzilakano wa siku. Ku lulamisiwa ka milayeni yimbirhi ku herise xiphepherhele hi ku helela.
- Xivumbeko xa ku tlhelela endzhaku hi ku miyela. Ku rhumeriwa eka xiteji swi humelerile 99.2% wa nkarhi, kambe LLM yi xiyile leswaku 31% wa ku rhumeriwa ka xiteji loku "humeleke" ku landzeriwe hi ku tirhisiwa kun'wana ka vukorhokeri lebyi fanaka ku nga si hela timinete ta 45 — leswi ringanyetaka leswaku ku rhumeriwa ko sungula ku onhiwile hi ndlela yo tirha hambileswi ku hundzeke swikambelo hinkwaswo. Leswi swi endle leswaku ku kumiwa leswaku xikambelo xa ku hlanganisiwa a xi hundza hikwalaho ka tinhlamulo leti hlayisiweke ku suka eka vukorhokeri bya mock.
- Xiphiqo xa Musumbhunuku nimixo. Minkarhi ya layini yo aka yi tlakuke hi 340% hi Musumbhunuku wun’wana na wun’wana exikarhi ka 9:00 na 10:30 AM hi nkarhi wa laha kaya, hikuva vatumbuluxi lava a va tirha hi mahelo ya vhiki hinkwavo va susumetele ku cinca ka vona va nga si yima. Ku lulamisiwa a ku nga ri ka xithekiniki — a ku tirha: ku hlamarisa xiyimiso xa CI runner pool scaling ku langutela ku tlakuka ka Musumbhunuku.
- Mujeko wa muhlengeleti lowu ku nga riki na munhu loyi a wu vekeke. 67% wa ku aka ka C++ a ku tirha handle ka ku hlengeletiwa loku engetelekeke loku pfuriweke, ku engetela xiringaniso xa timinete ta 3.8 hi ku aka. Mujeko a wu tsariwile eka nkongomiso wa ku nghena kambe a wu si tshama wu engeteriwa eka xifaniso xa vukorhokeri bya CI lebyi avelaniwa.
"Swihoxo swo durha swinene a hi leswi swi tshoveka xitirhisiwa xa wena. Hi swona leswi yivaka hi ku miyela 30 wa tisekoni eka ku aka kun'wana na kun'wana, siku na siku, ku ringana malembe — ku kondza munhu a hetelela a vutisa xivutiso lexi faneleke xa dataset leyinene."
Ku aka Leyara ya Vutlhari bya CI leyi Tirhaka
Xikambelo xi hi khorwisile leswaku nxopaxopo wa log lowu fambiwaka hi LLM a hi nchumu lowuntshwa — i vuswikoti bya ntiyiso bya ntirho. Kambe ku yi endla yi tirha swi lava vumaki lebyi anakanyisisiweke. A wu nge swi koti ku phayipha ntsena ti-raw log eka xihlanganisi xa mbulavurisano ivi u langutela tinhlamulo leti pfunaka. Sisiteme yi lava xivumbeko, naswona yi lava ku hlanganisiwa eka maendlelo ya ntirho lawa vanjhiniyara se va ma tirhisaka.
Hi tshamisekile eka endlelo ra swiyenge swinharhu. Xiyimo xo sungula i ku hlawula hi ku tisungulela: ku aka kun’wana na kun’wana loku tsandzekeke hi ku tisungulela ku kuma ku hlawuriwa hi xiyenge xa xivangelo xa rimitsu (xitirhisiwa, ku titshega, loji ya xikambelo, vukorhokeri, kumbe xiphepherhele) hi xikoro xa ku tshemba. Leswi ntsena swi hungute nkarhi wa xikarhi wo lulamisa ku tsandzeka ka ku aka hi 34%, hikuva vanjhiniyara a va nga ha boheki ku heta timinete ta khume va hlaya swiviko ntsena ku kuma laha va nga sungulaka ku languta kona. Xiyenge xa vumbirhi i ku kumiwa ka mikhuva: ku gayela ka vhiki na vhiki loku humesaka swivumbeko leswi humelelaka — ku engetela mimpimo ya ku tsandzeka, ku kula ka minkarhi yo aka, ku sayina swihoxo leswintshwa — swi nga si va swa nkoka. Xiyenge xa vunharhu i vulavisisi bya vuhlanganisi: xihlanganisi laha vanjhiniyara va nga vutisaka swivutiso swa ririmi ra ntumbuluko mayelana na matimu yo aka, ku fana na "Ha yini vukorhokeri bya X byi tsandzekile ngopfu endzhaku ka ku humesiwa ka Dzivamisoko?" kumbe "Hi xihi xivangelo lexi tolovelekeke xa swihoxo swa ku hela ka nkarhi eka phayiphi ya hakelo?"
Eka swipano leswi fambisaka matirhelo yo rharhangana — ngopfungopfu leswi lawulaka mintirho yo tala ya bindzu ku fana na CRM, ku endla ti-invoice, ku hakela, na vuxopaxopi hi ku tirhisa tipulatifomo to fana na Mewayz, leyi hlela 207 wa mimojula leyi hlanganisiweke — muxaka lowu wa ku langutiwa wu va wa nkoka swinene. Loko ku tirhisiwa kun’we ku khumba maendlelo ya ntirho lama langutaneke na vaxavi, loji ya ku hakela, na tisisiteme ta HR hi nkarhi wun’we, ku twisisa ku titshega eka phayiphi ya wena ya CI a hi ku hlawula. I swa nkoka ku hlayisa ku tshembheka loku 138,000+ wa vatirhisi va titshegeke ha kona.
💡 DID YOU KNOW?
Mewayz replaces 8+ business tools in one platform
CRM · Invoicing · HR · Projects · Booking · eCommerce · POS · Analytics. Free forever plan available.
Start Free →I Yini Lexi Nga Tirhiki (Sweswi)
Ku tshembeka i ka nkoka ku tlula hype. Ku ni swipimelo leswi nga erivaleni eka endlelo leri leswi un’wana ni un’wana loyi a ri kambisisaka a faneleke a swi twisisa. Ti-LLM ti vona swilo leswi nga vonakiki, naswona loko ti vona swilo leswi nga vonakiki hi ti- CI log, vuyelo bya kona byi nga ha va byi hoxile hi ndlela leyi khorwisaka. Hi vonile sisiteme hi ku tiyiseka yi vula leswaku ku tsandzeka ku aka ku vangiwa hi mpfilumpfilu wa ku titshega lowu nga si tshamaka wu va kona, lowu heleleke hi tinomboro ta vuhundzuluxeri lebyi endliweke. Phayiphi ya RAG yi hunguta leswi swinene, kambe a yi swi herisi. Vutivi byin’wana na byin’wana lebyi sisiteme yi byi humesaka bya ha lava ku tiyisisiwa hi vanhu ku nga si teka goza.
Xikalo xi tshama xi ri ntlhontlho. Loko sisiteme yo vuyisa yi nga khoma swivutiso hi ndlela leyinene, ku vekiwa ka swikombo swo sungula na ku nghenisa tilog letintshwa swa durha hi tlhelo ra xibalo. Hi lulamisa kwalomu ka 800,000 wa milayeni leyintshwa ya log siku na siku, naswona ku hlayisa index yi ri leyintshwa swi lava switirhisiwa leswi tinyiketeleke. Eka swipano leswintsongo, xibalo xa ku vuyeriwa ka ntsengo swi nga ha va swi nga tsakeli endlelo leri — hi xiringaniso a swi si tsakela. Loko tihakelo ta modele ti ya emahlweni ti ya ehansi (ti wile hi kwalomu ka 90% eka tinhweti ta 18 leti hundzeke eka vuswikoti byo ringana), ikhonomi yi ta cinca.
Ku tlhela ku va na xivutiso xa vuhlayiseki. Tilog ta CI ti nga va na swihundla — swilotlelo swa API, tintambhu ta vuhlanganisi, ti-URL ta le ndzeni — hambileswi ku nga na matshalatshala lamanene yo swi hlantswa. Ku rhumela datha leyi eka ti-API ta LLM ta le handle swi nghenisa khombo. Hi hunguta leswi hi phayiphi ya laha kaya yo hlantswa na hi ku fambisa inference eka timodeli leti ti rhurheleke eka vuhlayiselo lebyi nga na vuxiyaxiya, kambe swi engetela ku rharhangana na ntsengo. Swipano swi fanele ku kambisisa hi vukheta modele wa swona wa nxungeto swi nga si tirhisa nchumu wun’wana na wun’wana lowu fanaka.
Ku Sungula Handle ka Titerabyte
A wu lavi dataset leyikulu kumbe ntlawa wa vunjhiniyara bya ML lowu tinyiketeleke ku sungula ku humesa nkoka eka tilog ta wena ta CI. Hi leyi masungulo ya pragmatic lawa xipano xin’wana na xin’wana lexi nga na madzana ma nga ri mangani ya ku aka hi vhiki xi nga xi tirhisaka:
- Sungula hi ku hlawuriwa ka ku tsandzeka. Rhumela ehandle masiku ya wena yo hetelela ya 90 ya tilog ta ku aka leti tsandzekeke. Tirhisa API yin’wana na yin’wana ya LLM ku hlawula ku tsandzeka kun’wana na kun’wana hi swiyenge. Hambi ku ri taxonomy yo olova (infra vs. code vs. config vs. flake) yi nyika nkoka wa xihatla wa ku rhangisa emahlweni.
- Landzela mikhuva ya nkarhi wo aka. Hlawula switempe swa nkarhi ku suka eka tilog ta wena ku tumbuluxa nxaxamelo wa nkarhi wa nkarhi wo aka hi xiteji xa phayiphi. Ku phamela ti anomalies eka LLM leyi nga na mongo wa log lowu rhendzeleke na ku kombela ti root cause hypothesis.
- Endla leswaku swivutiso leswi "nga erivaleni" swi tirha hi ku tisungulela. Veka xihuku xa le ndzhaku ka ku tsandzeka lexi rhumelaka milayeni yo hetelela ya 500 ya ku aka loku tsandzekeke eka LLM hi xitsundzuxo: "Katsakatsa ku tsandzeka loku ka CI eka xivulwa xin'we ivi u ringanyeta ku lulamisiwa loku nga ha vaka kona swinene." Leswi ntsena swi hlayisa 5-10 wa timinete hi ku tsandzeka eka mininjhere un’wana na un’wana eka xipano.
- Aka akhavhiyu leyi laviwaka. Tirhisa ku nghenisiwa ku endla leswaku matimu ya wena ya log ya vutisisiwa hi ririmi ra ntumbuluko. Switirhisiwa swo fana na LangChain na LlamaIndex swi endla leswaku leswi swi fikeleleka hi ndlela yo hlamarisa, hambi ku ri eka swipano leswi nga riki na ntokoto wa ML.
Xilotlelo i ku sungula hi switsongo, u tiyisisa leswaku vutivi byi lulamile, naswona u ndlandlamuxa hakatsongo-tsongo. Ikhosisteme ya switirhisiwa swa muxaka lowu wa nxopaxopo yi vupfa hi xihatla, naswona leswi a swi lava switirhisiwa swa ntolovelo eka lembe leri nga hundza swi ya swi kumeka tanihi swiphemu leswi nga xavisiwiki.
Vumundzuku I Vutlhari bya Ntirho
Leswi hi vulavulaka hi swona hakunene a hi nxopaxopo wa log ntsena — i ku cinca ka xisekelo ku ya eka vutlhari bya matirhelo. Endlelo leri fanaka leri tirhaka eka tilog ta CI ri tirha eka mathikithi ya nseketelo wa vaxavi, datha ya tiphayiphi to xavisa, ku cincana ka swa timali, na maendlelo ya ntirho ya matirhelo. Thread leyi tolovelekeke hileswaku tinhlengeletano ti tumbuluxa nhlayo leyikulu ya datha ya matsalwa lama nga hlelekangiki ngopfu leyi nga na swivumbeko leswi nga tekiwaka goza, naswona ti-LLM ti faneleka hi ndlela yo hlawuleka ku kuma swivumbeko sweswo.
Leswi hi swona swi endlaka leswaku tipulatifomo leti hlanganisaka matirhelo ya mabindzu ti va na vuyelo bya xivumbeko. Loko datha ya wena ya CRM, vufambisi bya phurojeke, ku endla ti-invoice, tirhekhodo ta HR, na vuxopaxopi hinkwaswo swi hanya eka sisiteme yin’we — tanihilaha swi endlaka hakona eka swipano leswi tirhisaka xivumbeko xa modyuli leyi hlanganisiweke ya Mewayz — vuswikoti bya vutlhari byo tsemakanya domain bya andza. Xivumbeko eka tilog ta wena ta CI xi nga ha fambelana na ku churn ka vaxavi. Ku tlakuka ka mathikithi ya nseketelo ku nga ha vhumbha ku tsandzeka ka ku rhumeriwa. Swihlanganisi leswi swi vonaka ntsena loko datha yi tshama eka tisisiteme leti hlanganisiweke ku tlula tisilo leti nga toxe.
Swipano leswi nga ta humelela eka khume ra malembe lama taka a hi swona leswi nga na vanjhiniyara vo tala kumbe mipimanyeto leyikulu. Hi vona lava dyondzaka ku yingisela data ya vona — ku katsa na ti terabytes ta yona leti va nga tshama va ti lahla. Ti log ta wena ta CI ta vulavula. Xivutiso hi leswaku xana u lunghekele ku twa leswi va swi vulaka.
Swivutiso Leswi Vutisiwaka Nkarhi Na Nkarhi
Xana ti-LLM ti nga kuma tipheteni leti pfunaka hakunene eka tilog ta CI?
Hi ku hetiseka. Timodelo letikulu ta ririmi ti humelela eka ku kuma swivumbeko leswi vuyeleriwaka eka matsalwa lamakulu lama nga hlelekangiki. Loko ti kongomisiwa eka ti terabytes ta ti CI logs, tinga humesa ku yelana ka ku tsandzeka, ti flaky test signatures, na ku lwisana ka ku titshega loku vanjhiniyara va vanhu vanga ta ka vanga ku khomi hi voko. Xilotlelo i ku hlela phayiphi ya ku dya hi ndlela leyinene leswaku modele wu amukela swiphemu swa log leswi fuweke hi ndlela leyinene, leswi fuweke hi mongo ku tlula mpfumawulo lowu nga si swekiwaka.
Hi tihi tinxaka ta ku tsandzeka ka CI leti nga vhumbhiwaka hi ku tirhisa nxopaxopo wa log?
Nxopaxopo wa log lowu fambiwaka hi LLC wu nga vhumbha ku hela ka nkarhi loku fambelanaka na switirhisiwa, ku tsandzeka ka xiboho xa ku titshega loku vuyeleriwaka, ku wa ka ku aka loku bohiweke hi memori, na swikambelo swa flaky leswi hlohloteriwaka hi tindlela to karhi ta khodi. Yitlhela yi kombisa ti slow-creeping regressions laha minkarhi yo aka yi tlakukaka hakatsongotsongo eka mavhiki. Swipano leswi tirhisaka endlelo leri hi ntolovelo swi khoma swivumbeko swa ku tsandzeka ka cascading switsutsumi swimbirhi ku ya eka swinharhu swi nga si va swiendlakalo swo sivela eka ku tirhisiwa ka vuhumelerisi.
Xana u lava datha yo tanihi kwihi ya CI log loko nxopaxopo wu nga si va wa nkoka?
Tipheteni leti nga na nhlamuselo hi ntolovelo ti humelela endzhaku ko xopaxopa masiku ya 30 ku ya eka 90 ya matimu ya tiphayiphi lama yaka emahlweni ku tsemakanya marhavi yo tala. Tidathaseti letintsongo ti humesa vutivi bya xiyimo xa le henhla, kambe nkoka wa xiviri wu huma eka ku cross-referencing magidi ya ku tsutsuma ko aka. Eka swipano leswi lawulaka maendlelo ya ntirho yo rharhangana etlhelo ka tiphayiphi ta vona ta CI, tipulatifomo to fana na Mewayz ti nyika 207 wa mimojula leyi hlanganisiweke ku sukela eka $19/mo ku hlanganisa datha ya ntirho eka app.mewayz.com.
Xana ku phamela tilog ta CI eka LLM i khombo ra vuhlayiseki?
Swi nga va loko swi khomiwa hi vusopfa. Tilog ta CI ti tala ku va na swilo leswi cinca-cincaka swa mbango, swilotlelo swa API, ti-URL ta le ndzeni, na vuxokoxoko bya switirhisiwa. Loko u nga si tirhisa tilog hi ku tirhisa LLM yihi kumbe yihi, u fanele ku tirhisa tiphayiphi ta ku pfuxeta leti tiyeke leti hluvulaka swihundla, switifiketi, na vuxokoxoko lebyi tivekaka bya munhu. Ku tirhisiwa ka modele loku tiyimeleke kumbe loku nga eka ndhawu ku hunguta swinene ku paluxiwa loko ku pimanisiwa na ku rhumela tilog leti nga si swekiwaka eka makumu ya swibumabumelo leswi simekiweke eka papa ra vanhu va vunharhu.
Try Mewayz Free
All-in-one platform for CRM, invoicing, projects, HR & more. No credit card required.
Get more articles like this
Weekly business tips and product updates. Free forever.
You're subscribed!
Start managing your business smarter today
Join 6,205+ businesses. Free forever plan · No credit card required.
Ready to put this into practice?
Join 6,205+ businesses using Mewayz. Free forever plan — no credit card required.
Start Free Trial →Related articles
Hacker News
Arc Prize Foundation (YC W26) Is Hiring a Platform Engineer for ARC-AGI-4
Apr 17, 2026
Hacker News
Tesla tells HW3 owner to 'be patient' after 7 years of waiting for FSD
Apr 17, 2026
Hacker News
All 12 moonwalkers had "lunar hay fever" from dust smelling like gunpowder (2018)
Apr 17, 2026
Hacker News
NeoGeo AES+: SNK announces reissue of retro console without emulation
Apr 17, 2026
Hacker News
Show HN: Smol machines – subsecond coldstart, portable virtual machines
Apr 17, 2026
Hacker News
Random musings: 80s hardware, cyberdecks
Apr 17, 2026
Ready to take action?
Start your free Mewayz trial today
All-in-one business platform. No credit card required.
Start Free →14-day free trial · No credit card · Cancel anytime